You play this kind of game with "automation" and you get burned. Build, configuration and deployment management are hard software problems and they need to be solved by smart people and real tools, not checkboxes and file selectors.
You play this kind of game with "automation" and you get burned. Build, configuration and deployment management are hard software problems and they need to be solved by smart people and real tools, not checkboxes and file selectors.
You can't hack what you can't see, basically. And changes to the project file (or whatever other "wizard guided" IDE output you're using) are opaque by definition.
I am having a hard time truly understanding the difference (not that we use VS stuff for deployment anyways, but that is for other reasons).
A developer could add custom deployment info to the manual file and accidentally commit it just as easily.
You just do not like GUIs (which is a perfectly acceptable stance) but it has nothing to do with some fundamental difference.
Just because they supply a GUI doesn't mean it's the only way to change project settings. I'm not sure what led you to believe that.
While I'm probably in the minority, I edit my *.csproj files by hand and build from the command line for several of my projects. I do this for the exact reasons that you mention: I like control in the engineering process.
I'm unhappy that VS 2013 is putting user settings in the proj file willy nilly and I hope that it's fixed soon.
Your distinction is convenient and meaningless.
But even then, I need different settings for debug/release or x86/amd64 - and it's not really helping when there is incosistent naming across SDK's coming from Microsoft itself - where you can't match with one variable (say $(Target)) the given architecture, and it seems every group at Microsoft follows different conventions of how to name things.